Improved Recurrent Neural Networks for Session-based Recommendations
Y. Tan, X. Xu, and Y. Liu. Proceedings of the 1st Workshop on Deep Learning for Recommender Systems, page 17--22. New York, NY, USA, ACM, (2016)
DOI: 10.1145/2988450.2988452
Abstract
Recurrent neural networks (RNNs) were recently proposed for the session-based recommendation task. The models showed promising improvements over traditional recommendation approaches. In this work, we further study RNN-based models for session-based recommendations. We propose the application of two techniques to improve model performance, namely, data augmentation, and a method to account for shifts in the input data distribution. We also empirically study the use of generalised distillation, and a novel alternative model that directly predicts item embeddings. Experiments on the RecSys Challenge 2015 dataset demonstrate relative improvements of 12.8% and 14.8% over previously reported results on the Recall@20 and Mean Reciprocal Rank@20 metrics respectively.
%0 Conference Paper
%1 tan2016improved
%A Tan, Yong Kiam
%A Xu, Xinxing
%A Liu, Yong
%B Proceedings of the 1st Workshop on Deep Learning for Recommender Systems
%C New York, NY, USA
%D 2016
%I ACM
%K recommendation rnn session
%P 17--22
%R 10.1145/2988450.2988452
%T Improved Recurrent Neural Networks for Session-based Recommendations
%U http://doi.acm.org/10.1145/2988450.2988452
%X Recurrent neural networks (RNNs) were recently proposed for the session-based recommendation task. The models showed promising improvements over traditional recommendation approaches. In this work, we further study RNN-based models for session-based recommendations. We propose the application of two techniques to improve model performance, namely, data augmentation, and a method to account for shifts in the input data distribution. We also empirically study the use of generalised distillation, and a novel alternative model that directly predicts item embeddings. Experiments on the RecSys Challenge 2015 dataset demonstrate relative improvements of 12.8% and 14.8% over previously reported results on the Recall@20 and Mean Reciprocal Rank@20 metrics respectively.
%@ 978-1-4503-4795-2
@inproceedings{tan2016improved,
abstract = {Recurrent neural networks (RNNs) were recently proposed for the session-based recommendation task. The models showed promising improvements over traditional recommendation approaches. In this work, we further study RNN-based models for session-based recommendations. We propose the application of two techniques to improve model performance, namely, data augmentation, and a method to account for shifts in the input data distribution. We also empirically study the use of generalised distillation, and a novel alternative model that directly predicts item embeddings. Experiments on the RecSys Challenge 2015 dataset demonstrate relative improvements of 12.8% and 14.8% over previously reported results on the Recall@20 and Mean Reciprocal Rank@20 metrics respectively.},
acmid = {2988452},
added-at = {2017-02-26T18:07:29.000+0100},
address = {New York, NY, USA},
author = {Tan, Yong Kiam and Xu, Xinxing and Liu, Yong},
biburl = {https://www.bibsonomy.org/bibtex/297a26821c78b1b68b26ca559bf9a6c2e/nosebrain},
booktitle = {Proceedings of the 1st Workshop on Deep Learning for Recommender Systems},
doi = {10.1145/2988450.2988452},
interhash = {7dd7bde2439a64bdf01bac1202d701fd},
intrahash = {97a26821c78b1b68b26ca559bf9a6c2e},
isbn = {978-1-4503-4795-2},
keywords = {recommendation rnn session},
location = {Boston, MA, USA},
numpages = {6},
pages = {17--22},
publisher = {ACM},
series = {DLRS 2016},
timestamp = {2017-02-26T18:07:29.000+0100},
title = {Improved Recurrent Neural Networks for Session-based Recommendations},
url = {http://doi.acm.org/10.1145/2988450.2988452},
year = 2016
}